Loading research records…
HYSONLAB · SEQUENCE → STRUCTURE → EVIDENCE
Small molecules.
Deeper understanding.
Explore published antimicrobial peptides, inspect their molecular
structures, and discover neighbors in sequence and structure space.
—DRAMP records
—Experimental chains
—Source SMILES
—Static dataset · Under 1 GB
Peptide explorer DRAMP + RCSB PDB
↓ Export CSV⌕
PEPTIDESLENGTH
CALCULATE & FIND PUBLISHED NEIGHBORS
Sequence workbench.
Explore charge, hydropathy, mass, and amphipathicity. Find related sequences and their original studies.
1–500 canonical residues · Free N terminus · No disulfide correction
Calculated descriptors are sequence-based approximations, not activity predictions. Neighbors require a source study reference and are deduplicated by exact sequence. Descriptor search compares all records at pH 7 with documented fixed scales. Source chemical modifications may make calculated values differ from the actual molecule.
DESIGN ON THE WHEEL → READ THE SEQUENCE
Helical-wheel designer.
Choose a peptide length. Fill the numbered wheel positions to build a linear N→C sequence.
Design any canonical peptide; search the published DRAMP collection for matching positions.
1–72 residues · 100° per residue · residues after position 18 use additional display rings to avoid overlap. Ring layout is a projection aid; it does not describe a cyclic peptide or predict a folded structure.
Click a position to edit
Blue: basic · Coral: acidic · Olive: hydrophobic · Gray: other
Keyboard: focus a numbered position and type an amino acid. ←/→ selects the previous/next sequence position; Delete clears it. Letters advance in N→C order.
Position 1
Choose a residue, or leave this position open as a wildcard.
Add a residue to see live matches.
? marks an unspecified residue. Sequence order follows position numbers, not clockwise order around the circle.Live search & design properties
Filter matching peptides by calculated properties
Candidate properties use source-terminal annotations at pH 7, with the existing sequence-only assumptions. Filters are not activity or safety predictions.
Complete sequences can be analyzed and searched by similarity in the existing workbench. Partial-wheel search requires at least one specified residue, exact sequence length, and a source study reference. These tools design sequences; they do not establish antimicrobial activity or helicity.
COMPARE EXPERIMENTAL GEOMETRY
Structure neighbors.
Fit observed peptide backbones and inspect their differences in three dimensions.
8–150 observed canonical residues. Global sequence-guided alignment + Kabsch Cα superposition, ≥80% coverage of the longer chain. RMSD ranks geometry after that alignment; this is not an independent fold-search engine. NMR conformations and experimental environments affect the result.
TRACEABLE DATA, TRANSPARENT CALCULATIONS
Data & methods.
Official source annotations, local experimental structures, and reproducible calculations.
DRAMP collection
↓ Complete SQLite database · Source code ↗
Loading provenance…
Activity, target organisms, hemolysis, cytotoxicity, modifications and references are preserved as source text, not converted into normalized assay measurements. Predicted and measured properties are not mixed.
Official downloads ↗Molecular visualization
3Dmol.js renders static all-atom PDB files with cartoon, sticks, spheres and molecular surfaces. These are experimental first-model coordinates; missing atoms and modified residues follow the deposited file.
PDB chains are linked only when their observed canonical sequence exactly matches a DRAMP sequence or a contiguous fragment covering ≥80% of it. Structure overlays use observed Cα atoms only.
3Dmol.js ↗ · RCSB PDB ↗Descriptors & similarity
Mass: average residues + water, optionally C-amidated. Charge/pI: Henderson–Hasselbalch. GRAVY and hydrophobic moment use Kyte–Doolittle; the moment and wheel assume an ideal 100° α-helix, not a predicted structure. Extinction assumes reduced cysteines.
Sequence score: 100 × (1 − global edit distance / longer length). Alignment identity includes gap columns in the denominator. Property score: 100 / (1 + scaled RMS descriptor difference). Neither score establishes antimicrobial function.
DRAMP 5.0 reference ↗Dataset & browser manifest
Sequence and uploaded PDB calculations run on your device. GitHub serves public data and may log ordinary page/file requests. External study links open their respective providers.